Faster Re-Check

[joshace]

The way the re-check works now is that every bit of the file is verified to determine which parts are complete and which are not.  That is necessary on an incomplete file.

HOWEVER, on a complete file, it should do a rapid analysis of the file on the drive, and come up with a checksum.  If the checksum is correct, then that file should be assumed to be 100% correct without verifying all of it.  

The checksum on an incomplete file is not going to be correct.  

That would speed things up QUITE a bit.. because most of the rechecks are of completed files.  

 

>> In other words.. if the checksum is correct, skip the thorough check.  The odds of having the correct checksum when the files are not identical are infinitesimal. 

[DreadWingKnight]

per-file checksums are not stored.

And what is your net.low_cpu value set to?

[joshace]

my net.low_cpu is set to FALSE

 

I have extremely fast internet..  and my downloads are often 3m to 7m per second.. and uploads over 1m

 

[joshace]

15 hours ago, DreadWingKnight said:

per-file checksums are not stored.

 

Hmm..  well.. that brings up another idea... how about adding that feature to Utorrent.. where it DOES store the per file checksum when the file is first completed... Then in the future.. when rechecks are done... if the checksum exists, then recalculate the checksum from the file that is selected to "re-check", and if the recalculated checksum matches the original checksum.. then skip the long thorough verification of every bit of the file?